Transitions to Democracy and

the Rise of Nationalist Conflict

T he centerpiece of American foreign policy in the 1990s was the claim

that promoting the spread of democracy would also promote peace.
Noting that no two democracies have ever fought a war against each
other, President Bill Clinton argued that support for democratization
would be an antidote to international war and civil strife. 1 Yet paradoxi­
cally, the 1990s turned out to be a decade of both democratization and
chronic nationalist conflict.
While the world would undoubtedly be more peaceful if all states
became mature democracies, Clinton's conventional wisdom failed t<l
anticipate the dangers of getting from here to there. Rocky transitions to
democracy often give rise to warlike nationalism and violent ethnic COr1­

1 See his 1994 State of the Union address, "Transcript of Clinton's Address,» Nnu

York Time5, January 26, 1994, Al7. On the absence ofwar between democracies, see
Michael Doyle, "Liberalism and World Policics," American Political Science Rrvinu
80:4 (December 1986), 1151-1169.

15
16 From w,ting to Violence
flicts. Since the French Revolution, the earliest phases of democratization
have triggered some of the world's bloodiest nationalist struggles.
Spreading the benefits of democracy worldwide is a wonhy long-run
goal. However, strategies for accomplishing this must be guided by a real­
istic understanding of the politics of the transition. Naively pressuring
ethnically divided authoritarian stares to hold instant elections can lead to
disastrous results. For example, international financial donors forced free
and fair elections on the leaders of the small central Mrican country of
Burundi in 1993, and within a year some 50,000 Hutu and Tutsi were
killed in ethnic strife there. And yet many other democratic transitions
succeed without triggering nationalist violence. Understanding the condi­
tions that permit such successful transitions should be the first step
toward designing policies to pave the way toward democracy. To that end,
this book explains why democratization often causes nationalist conflict,
and why it sometimes does not. Drawing on that analysis, I prescribe
ways to make democratic transitions less dangerous.
Liberal Optimism Confronts the Nationalist Revival
of the 19908
When the Berlin Wall came down in 1989, a euphoric vision briefly cap­
tured the American imagination. Liberalism had triumphed over its two
ideological competitors in the twentieth century, communism and fascist
nationalism, and no new challengers were in sight. Empires and dictator­
ships were collapsing. Democratization was sweeping formerly authoritar­
ian countries in Latin America, southern Europe, and Eastern Europe,
and even making inroads in East Asia. Vinually everywhere states were
adopting market economies. Global economic interdependence was con­
tinuing to deepen. Liberal, American-based mass news media and popu­
lar culture were achieving global reach. This victory of liberalism, it was
claimed, would usher in "the end ofhistory."2
Believing that all good things go together, liberal commentators argued
that war was becoming obsolete, at least among the liberalizing countries
, Francis Fukuyama, "The End of History," The National Interest 16 (Summer
1989),3-18.
Transitions to Democracy and the Rise ofNationalist Conflict "17
that were establishing the dominant global trend. 3 A learned tome pub­
lished in 1990 concluded that nationalism, commonly defined as the
doctrine that each cultural group should have its own state, was rapidly
heading into the dustbin of history because states organized around single
nations could no longer cope with an increasingly interdependent, global­
izing world.' Residual snunbling blocks in the path of triumphant liberal­
ism could be overcome with the help of an energetic ser of international
institutions-United Nations troops to keep the peace, and International
Monetary Fund experts to lure countries into the liberal fold and to
enforce the rules of fiscal prudence. In keeping with this vision, President
Clinton explained that promoting democratization would be a watch­
word of U.S. foreign policy-because democracies never fight wars
against each other, they trade freely with each other, and they respect the
human rights of their citizens. s
This vision tarnished quickly. War has been endemic since the collapse
of the Berlin Wall. Nor have these been trivial wars at the periphery of the
international order: the world's oil supply was at risk in the 1991 Gulf
War; in June 1991 the Yugoslav army battled Slovenian separatists scarcely
a hundred miles from Vienna, and NATO's air forces mounted a sustained
bombing campaign throughout Serbia during the 1999 Kosovo conflict.
Nationalist rhetoric, far from being discredited, came back into vogue. A
quarter of the electorate in Russia's fledgling democracy voted in 1993 for
the party of a neofuscist anti-Semite, Vladimir Zhirinovsky (only a third
voted for Hitler in 1932). In civil wars from Somalia to Bosnia, the armed
forces of the liberal international community were bedeviled, attacked,
and held hostage by local thugs. Ethnic mayhem in 1994 caused over half
a million deaths in Rwanda, after Belgian peacekeepers were killed on the
first day of the genocide campaign against the T utsi minority.
As a result, the conventional wisdom was soon turned on its head: The
Atlantic Monthly relabeled the post-Communist world as "the coming anar-
Mueller, Retreatfrom Doomsday: The Obsolescence ofMajor \Wtr (New York:
Basic Books, 1989).
• E. J. Hobsbawm, Nations and Nationalism since 1780 (Cambtidge: Cambridge
University Press, 1990), Chapter 6.
\ In addition to the 1994 State of the Union address, see also the article
Clinton's national security adviser, Anthony Lake, "The Reach of Democtacy," lVew
York Times, September 23, 1994, A35.
18 From voting to Violence
chy;" and the eminent Harvard policical sciencist Samuel Huncington
announced that the furure would hold in store a "clash of dvilizacions."6 In
the view of such pWldits, cultural confliCts, whether along the fault lines of
whole civilizacions or simply among intermingled ethnic groups, would
become the defining cleavages of internacional relations in the coming era.
News media and policicalleaders commonly attributed these grim develop­
ments to "ancient hatreds" between inimical cultures, simmering fur cen­
turies and boiling over as soon as the lid of the cold war was lifi:ed off the
pot. This accoWlt was simple, intuicive, and reinforced daily by the justifica­
cions offeted by perpetrators of ethnic slaughter. For Western policidans
looking fur an easy excuse to limit their involvement in W1Seemly struggies,
the story of ancient hatreds also had the advantage of portraying these dis­
putes as hopelessly intraCtable. But even those who retained the vision of
spreading liberal democracy to Wlaccustomed comers of the globe consid­
ered age-old ethnic pre,judices to be liberalism's major foe. President Clinton,
in his 1993 presidential inauguracion speech, remarked that "a generacion
raised in the shadows of the cold war assumes new responsibilicies in a world
warmed by the SWlShine of freedom but threatened still by ancient hatreds."1
The good news is that this view is largely incorrect. Most of the globe's
recent strife is not due to ancient cultural hatreds. In some cases, the war­
ring groups had experienced no armed conflict uncil relacively recently.
Serbs and Croats, for example, never fuught each other unci1 the twencieth
cenrury, and then largely because the Nazis installed an unrepresentacive
regime ofmurderers in Zagreb. 8 In other cases, occasional confliCts between
cultural or ethnic groups have been interspersed with long interludes of
6 Robert D. Kaplan, "The Coming Anarchy," The Atlantic MonthlJ 273:2
(February 1994), 44-76; Samud Huntington, "The Clash of Civilizations,"
Affairs 72:3 (Summer 1993),22-49.
7 QUOted in Susanne Hoeber Rudolph and Uoyd I. Rudolph, "Modem Hate,»
The New &public 208:12 (March 22, 1993),24. Clinton later acknowledged that his
"ancient hatreds" remarks were erroneous. See Katharine Seelye, "Clinton Blames
Milosevic, Not Fate, for Bloodshed," New l1Irk Times. May 14, 1999, A12.
S The first armed conflict pitting Serbs against Croats was World War I, when half
of the Austto-Hungaxian forces invading Serbia were Croats. However, a quarter of
this invasion force were Serbs from the border districts of Croatia, a group whose tra­
ditional occupation had been frontier guards against the Ottomans. Since Serbs
fOught on both sides, World War I can hardly be classified as an example of an
ancient ethnic con/lict. Dimitrije Djordjevic, "The Yugoslav Phenomenon," in
Transitions to Democracy and the Rise ofNationalist Conflict 19
amicable relacions, therefore cultural differences cannOt in themselves be a
sufficient explanation for the recent fighting. Few serious scholars attribute
nacionalism and ethnic conflict primarily to ancient cultural hatreds. 9
The bad news, however, harbors a deep irony: the very trends that lib­
erals saw as bringing the end of history have in many instances fueled the
revival of nationalism. The end of the authoritarian Soviet empire spurred
the aspiring leaders of many of its intermingled nations to establish their
own national states, whose conflicting claims to sovereignty and territory
often gave rise to disputes. 10 Elections often sharpened these ethnic and
national differences. Nationalist demagogues exploited the increased free­
dom of the press in some newly democracizing states to hijack public
debate for illiberal ends.n Painful adjustments to a market economy and
to international interdependence provided further opportunicies for
nationalist policidans who promised protection in a strong state, or who
distributed a shrinking economic pie along ethnic lines. At the same cime,
the globalization of media and culture often repelled rather than attracted
those who failed to prosper in a Westernized world. 12 Moreover, as some
critics argue, international organizacions sometimes caused more conflict
than they averted with their inept strategies of peacekeeping and their
strict philosophies of economic refurm. 13
Though surprising to liberal optimists, these developments of the
Joseph Hdd, The Columbia History ofFAstml Europe in the Twentieth Cmeury (New
York: Columbia University Press, 1992), 310.
, Although ethnic conflicts are a prevalent feature of the post-<:old war landscape,
they do not necessarily represent an increasing proportion of all conflicts. Myron
Weiner, "Bad Neighbors, Bad Neighborhoods: An Inquiry into the Causes of Refugee
Flows," International StCUrity 21:1 (Summer 19%), 21-22, points out that between
1%9 and 1992 the proportion of nonethnic refugee-causing conflicts increased.
10 Placing the Soviet collapse in comparative perspective is Karen Barkey and Mark
Von Hagen, eds., After Empire: Multiethnic Societies and Nation-Building (Boulder,
CO: Westview, 1997).
H Jack Snyder and Karen Ballentine, "Nationalism and the Marketplace of Ideas, »
International Security 21:2 (Fall 19%), 5-40.
11 Huntington, "The Clash of Civilizations,» 25-62; Mark Juergensmeyer, The
New Cold ~r." &ligious Nationalism Confronts the Secular State (Berkdey: University
of California Press, 1993).
13 Richard K. Betts, "The Ddusion ofImparrialIntervention," Foreign Affairs 73:6
(NovemberlDecember 1994),20-33; Susan Woodward, Balkan Tragedy (Washington,
DC: Brookings, 1995), Chapter 3. '
20 From Voting to Violence
1990s actually echoed long-standing patterns in the history of national­
ism, which I explore in subsequent chapters. Far from being an outmoded
throwback, nationalism is largely a reaction to the social changes of the
modern era. Western Europe went through these changes between the
French Revolution and the Second World War, an age that saw the rise of
modern nationalism and of popular warfare. During that period, democ­
ratization, economic development, and a revolution in the means of com­
munication fueled nationalism, which often took a militant form. States
being dragged by social change into a transition to democracy have been
more likely to participate in wars and more likely to start them than have
states whose regimes did not change. 14 The end of the cold war increased
the prevalence of nationalism by unleashing this dangerous transition
toward democratic, market societies in the post-Communist states.
Though democratization heightens a state's risk of war, historical evi­
dence shows that three our of four democratizing States nonetheless
avoided war in the decade after their democratization. Moreover, once
democracy became entrenched, no mature democracies have ever
fought wars against each other. In those countries where transitions to
democracy were fully consolidated during the 1990s, the rights of ethnic
minorities tended to improve, and ethnic conflicts were rare. 15
The central message of this book, consequently, constitutes a paradox.
On the one hand, the successful unfolding of a global, liberal-democratic
revolution might eventually undergird a more peaceful era in world poli­
tics. On the other hand, the transition to democratic politics is meanwhile
creating fertile conditions for nationalism and ethnic conflict, which not
only raises the COSts of the transition but may also redirect popular politi­
cal participation into a lengthy antidemocratic detour. The three most
nearly successful attempts to overturn the global balance of power through
aggression-those of Napoleonic France in 1803-15, Kaiser Wilhelm's
Germany in 1914-18, and Adolf Hitler's Germany in 1939-45-alJ came
l4 Edward D. Mansfield and Jack Snyder, «Democratization and the Danger of
War," International Security 20:1 (Summer 1995), 5-38; Mansfield and Snyder,
«Democratization and War,n Foreign Affairs 74:3 (May/June 1995),79-97.
15 Ted Robert Gurr, Peoples versus States: Ethnopolitical Conflict and
Accommodation at the End of the 20th Century (Washington, DC: US Instirute of
Peace, forthcoming); see also Ted Robert Gurr and Will H. Moore, "Ethnopolitical
Rebellion: A Cross-Sectional Analysis of (he 1980s with Risk Assessments for the
1990s," American Journal ofPoiiticai Science 41:4 (October 1997), 1079-1103.
Transitions to Democracy and the Rise ofNationalist Conflict 21
on the heels of failed attempts to democratize. Popular nationalistic ener­
gies, unleashed and perverted by the miscarriage of democratic reforms,
created the conditions that made possible these bids for global hegemony.
Thus, the process of democratization can be one of its own worst enemies,
and its promise of peace is clouded with the danger of war.
To promote democratization without heeding these risks would be
self-indulgent idealism. Yet to try to hold back the global social processes
that may stimulate nationalism-including demands for increased mass
participation in politics, the collapse of outworn empires, and the global­
ization of the economy and communications-would be equally unrealis­
tic. Instead, one of the key rasks for the international community will be
to distinguish the circumstances that make for a safe transition to liberal
democracy from those that lead to backlash, nationalism, and war. Insofar
as astute policy choices can help to create the more favorable conditi<>ns,
understanding the pitfalls of democratization is the first step toward
avoiding them. This is an analytical task for everyone who is engaged
with the seminal issues of our day: political leaders in the advanced
democratic states and in transitional states, journalists, human rights
activists, scholars, citizens, and even nationalists themselves. insofar as
they want to avoid costly missteps in the pursuit of their nations' goals.
In this introductory chapter, I will first define what I mean by the
terms nationalism and democratization; second, summarize the evidence
that nationalist conflict correlates with the early phases of democratiza­
tion; third, challenge the view that ancient popular rivalries explain rllls
correlation; fourth, briefly sketch my own alternative explanation, wh.ich
highlights the role of persuasion by nationalist elites; fifth, discuss the
practical implications of this theory for policy choices; and finally, pro­
vide a road map to the historical and contemporary case studies that con­
stitute the bulk of the book.
What Are Nationalism and Democratization?
Nationalism
In everyday usage, a variety of very different phenomena may all be
labeled :15 manifestations of nationalism, including ethnic riots, aggressive
foreign policies of fascist states, patriotism in democracies, and the peace­
22 From y&ting to Violence
Transitions to Democracy and the Rise ofNationalist Conflict 23

ful seeking of special rights for cultural groups. To avoid confusion, social
view of neighboring nation-states as inferior, hostile, and deserving of
scientists typically like to define terms more narrowly and precisely than
domination.'s Everyday usage assumes that these broader meanings are an
is common in everyday discourse.
integral part of the thing people call nationalism. I will try to show that
The most widely used scholarly definition of nationalism is by Ernest
common parlance links these phenomena not out of confusion but
Gellner, who defines nationalism as the doctrine that the political unit
because they have related causes, dynamics, and consequences, which a
(the state) and the cultural unit (the nation) should be congruent. '6
theory of nationalism and nationalist conflict ought to try to capture. To
According to this view, nationalism holds that the state, which is the
accommodate this, Gellner's definition, though a useful starting place,
organization that exercises sovereign authority over a given territory,
needs to be broadened.
should rule on behalf of a particular nation, defined as a group of people
I define nationalism, therefore, as the doctrine that a people who see
who feel they share a common culture. This formulation is theoretically
themselves as distinct in their culture, history, institutions, or principles
clear and historically plausible. Many self-proclaimed nationalist move­
should rule themselves in a political system that expresses and protect:S
ments have had as their central aim the acquisition of a state by a cultur­
those distinctive characteristics. 19 A nation is, therefore, a group of people
ally distinct people (like Croatian nationalists did in 1991), the inclusion
who see themselves as distinct in these terms and who aspire to self-rule.
of cultural brethren in an existing state (Hungarian nationalists in the
Nationalist conflict is defined as organized, large-scale violence motivated
1930s), or the domination of an existing state by a single cultural group
or justified by a nationalist doctrine.
(Estonian nationalists in the 1990s). Indeed, in an age when conflicts
By this definition, not all ethnic groups are nations; nor are all nations
rooted in ethnic nationalism are such a dominant concern, it is tempting
ethnic groups. There are many peoples who consider themselves to be
to highlight cultural distinctions in the very definition of nationalism.
culturally or historically distinct, for example the Cajuns of Louisiana,
Yet this definition would seem to leave out phenomena that common
but who lack a doctrine that claims a right to self-rule for the group.
usage, including usage by self--described nationalists, normally calls nation­
Based on wide-ranging historical research, Anthony Smith distinguishes
alism. For example, defining nationalism strictly in terms of shared culture
between an ethnic group, or ethnie (which has a distinctive consciousness
would seem to exclude militant loyalty to a state's political institutions or
based on a common language or culture, myths of common ancestry, or a
other principles not based on culture, such as the universalistic principles
common historical experience), and a nation (which seeks self-rule for
embodied in the U.S. Constitution. Similarly, defining the aim of nation­
such a groUp).20 Ethnic conflict involves nationalism only when a goal of
alism as achieving a sovereign state would seem to exclude the seeking of
the conflict is to establish or protect self-rule by the ethnic groUp.21
political rights shott of sovereign statehood by cultural groups, such as
those Quebecois nationalists who seek a furm of autonomy within the
Canadian federal state. Moreover, nationalists often do not stop at getting IS John Mearsheimer, "Back to the Future," International Security 15:1 (Summer
their own state. They frequently attempt to enshrine distinct cultural val­ 1990),5-56, calls this "hypernationalism:
l~ Snyder and Ballentine, "Nationalism and the Marketplace ofIdeas," 9-10.
ues in that state, discriminate in favor of coethnics living within its bor­
2. Anthony D. Smith, The Ethnic Origins of Nations (Oxford: Blackwell, 1986),
ders, try to incorporate ethnic brethren and historic national territories
Chaprers 1 and 2.
into the state, and militantly guard against encroachments by historic ene­ 21 The mosr ambitious expression of the demand for self-rule is the achievement of
mies of the nationY In some cases, the nation-state adopts a "nationalistic" sovereign statehood, whereby the nation establishes a bureaucratic apparatus that
monopolizes the legitimate use of force within the rerritory it governs (Gellner,
Nations, 3-5). Some nationalist movemenrs, however, seek more limited forms of
16 Ernesr Gellner, Nations and Narionalism (Ithaca: Cornell University Press,
self-rule, involving autonomous authority for the nation over a narrower range of
1983), 1-7.
issues such as educational and language policy, a separate legal code, taxation, or con­
17 Rogers Brubaker, Nationalism Reftamed (Cambridge: Cambridge University
trol over local police forces. The Quebecois political movement, tor example, wonld
Press, 1996), Chapters 3 and 4.
remain narionalist by my definition if ir sought to expand or prorecr the scope of self­
24 From Witing to Violence
Although nationalist doctrine derives political authority from the right of
a distinct people to rule themselves, nationalists do nor necessarily hold that
legitimate political processes require democratic voting. Rather, the right to
self-rule means that the national group should not be ruled by an alien peo­
ple or alien institutions. It also means that the nation's rulers, no matter how
they are chosen, must justify their policies in terms of the welfare, security,
and fulfillment of the national aims of the sovereign people. This ambiguity
between rule by the people and rule in the name of a people constitutes one
of the main attractions of nationalist doctrine to elites who seek to rule
undemocratically in an era of rising demands for a mass role in politics.
Nations may distinguish themselves from each other not only on the
basis of distinctive cultural traditions, but also on the basis of distinctive
political traditions, political institutions, and political principles..u Thus,
scholars commonly divide nationalisms into two types, ethnic and civic,
based on the nature of their appeals to the collective good and on their
criteria for including members in the group.23 Ethnic nationalisms, like
those of the Germans and the Serbs, base their legitimacy on common
culture, language, religion, shared historical experience, and/or the myth
of shared kinship, and they use these criteria to include or exclude mem­
bers from the national groUp.24 For example, German law offered citizen­
ship to people of German ancestry who reside in Russia, while it denied it
to many Turks who have lived in Germany all their lives. Civic nation­
alisms, like those of the British, the United States, and for the most part
the French, base their appeals on loyalty to a set of political ideas and
institutions that are perceived as just and effective. Inclusion in the group
depends primarily on birth or long-term residence within the nation's ter­
ritory, though sufficient knowledge of the nation's language and institu­
tions to participate in the nation's civic life may be a criterion for the
naturalization of resident aliens.
rule of the distinct Quebecois nation, even while remaining within the sovereign state
of Canada.
II Indeed, given Gellner's extremely broad definition of culture as "a system of
ideas and signs and associations and ways of behaving and communicating.» a
nation's political ideas and habits would count for him as culture. Gellner, Nations, 7.
13 Liah Greenfeld, Nationalism: Five Roads to Modernity (Cambridge: Harvard
University Press, 1992); Rogers Brubaker, Citizenship and Nationhood in France and
Germany (Cambridge: Harvard University Press, 1992).
" Smith, Ethnic Origins, 22-31.
Transitions to Democracy and the Rise ofNationalist Conflict 25
This distinction between civic and ethnic nationalism is especially cru­
cial in countries like contemporary Ukraine, where ethnic Russians and
Ukrainians live intermingled. In those conditions, basing political loyal­
ties on cultural or linguistic differences would be intensely divisive.
Consequently, Ukraine's leaders have for the most part prudently pro­
moted a civic-territorial form of nationalloyalry.
These categories are ideal types: no actual nation is purely civic or
purely ethnic. Ethnic groups that seek political goals normally set up
administrative institutions that function at least partly according to legal
criteria, not just cultural norms. Conversely, civic states are often built on
some discernible ethnic core, and over time, civic nations generate their
own civic culture and shared historical myths.25 Nonetheless, nations can
be placed on a continuum between the civic and ethnic ideal types
depending on whether loyalty to and inclusion in them is based primarily
on institutions or on culture. A definition of nationalism that is broa.d,
yet distinguishes between ethnic and civic variants, permits the investiga­
tion of the causes and consequences of both types. 26
In short, this definition of nationalism highlights popular self-rule as a
universal goal of nationalists but avoids smuggling democracy into the
very definition of nationalism. It also allows the exploration of the cul­
tural basis of political loyalty but avoids the mistake of equating national­
ism with ethnicity. Thus, it features some of the elements of nationalism
that are central to understanding the causes of nationalism and its conse­
quences for violent conflict.
Democratiuaion
The term democratization distinguishes between mature democracies and
democratizing states. In mature democracies, government policy. includ­
" Ibid., 153-54.
211 Although this definition of nationalism is broad, it does not subswne all types
of political loyalty. There are many doctrines of supreme political loyalty that a.re
based on alternative criteria for alignment and enmity, other than nations--e.g.,
social classes; smaller-scale kinship groups such as clans; local communities of indi­
viduals with face-to-face relationships; transnational religions; statUS groups such as
the aristocracy; professional organizations such as guilds and knightly orders; family
dynasties, empires, networks of patronage and personal reciprocity; or organizations
that stand for universalistic moral principles. '
26 From Voting to VioLmce
Transitions to Democracy and the Rise ofNationalist Conflitt 27

ing foreign and military policy, is made by officials chosen through free,
At what moment does a successfully democratizing state become a
fillr, and periodic elections in which a substantial proportion of the adult
mature democracy? When can its democracy be termed consolidated?
population can vote; the actions of officials are constrained by constitu­
Some scholars use the "two turnover rule" to define democratic consolida­
tional provisions and commitments to civil liberties; and government can­
tion: that is, a democracy is consideted consolidated when power has
didates sometimes lose elections and leave office when they do. Freedom
changed hands twice as a result of free and fair elections. Others say that
of speech, freedom to organize groups to COntest elections, and reasonably
democracy is consolidated when it is "the only game in town": that is,
equitable representation of varied viewpoints in the media are presumed to
when no significant political party or group seeks to come to power by
be preconditions for a free and fair election. I define states as democratiz­
means other than winning a free and fillr election. 29 Finally, others mea­
ing if they have recently adopted one or more of these democratic charac­
sure the degree to which the country has achieved the institutional and
teristics, even if they retain important nondemocratic features. 27
legal characteristics of a mature democracy, using indicatots such as com­
The category of democratizing states is a very broad one. It includes
petitive politics, regular elections, broad participation, constraints on
states like the Czech Republic in the early 1990s, which made a transition
arbitrary use of executive power, free speech, and respect for civil liberties,
from complete autocracy to virtually complete democracy. However, it
including minority rights. When a country achieves a high enough score
also includes the former Yugoslavia just before its breakup in 1991, when
on almost all of these dimensions, it is said to have consolidated its
elections were contested for the first time in circumstances of somewhat
democracy.3o States that have crossed this line by any of the above criteria
freer speech, yet electoral fairness and the rule of law were hardly well
are mature democracies, no longer democratizing states. 31
established.2l!

27 Snyder and Ballentine, "Nationalism and the Marketplace of Ideas." 6 (fri. 5). The Link. between Democratization and Nationalist Conffict:

For the purpose of a quantitative study of the relationship between democratization Some Evidence

and war. the findings of which I occasionally invoke in this book, Edward Mansfield
and I identified specific thresholds for several types of institutional change from an
Diverse evidence points to a connection between democratization and
autocracy to a mixed regime and ttom a mixed regime to a democracy. We labeled
states crossing anyone of these thresholds as democratizing. See Edward D.
conflicts fueled by nationalism. As more people begin to playa larger role
Mansfield and Jack Snyder. "Democratization and the Danger of War," 8-10. We in politics, ethnic conflict within a country becomes more likely, as does
relied on the categories and case codings of Ted Gurr's Polity II and III databases. international aggression justified by nationalist ideas,
However. in narrative discussions of cases in this book, I do not rely solely on these
particular quantitative thresholds. Rather, I describe qualitatively the specific changes
Nazis in the Weimar Republic; in democratizing states, such as the German Navy
in institutions or civil liberties that warrant the designation of democratization.
League before World War I; and in states untouched by significant democratization,
2lI Though democratization is a broad term, scholars sometimes employ a still
such as the nationalist Chinese Kuomintang movement of Chiang Kai-shek.
broader term-that is, increased political participation. Political participation may
20 See Juan Linz and Alfred Stepan, Probkms of Democratic Transition and
take democratic forms, such as voting or joining a political party, but it may also rake
ConsoLidation (Baltimore: Johns Hopkins University Press, 1996), Chapter 1.
nondemocratic forms, such as rioting, striking to express political demands, or form­
,. Both Gu.rr and Freedom House aggregate scores across multiple dimensions in
ing mass-based paramilitary movements. See Samuel Huntington, PoLiticaL Order in ff
this way. See Mansfield and Snyder, "Democratization and the Danger ofWar and
Changing Societies (New Haven: Yale University Press, 1968). Where these latter
"Democratic Transitions, Institutional Strength, and War" (manuscript, 1999), for a
forms of political participation are prevalent, democratic elections are not "the only
discussion of the precise cutoff poinrs used to distinguish mixed regimes from marure
game in town," and perhaps not even the main game. Historically, one important
derp.ocracies.
form of increased mass political participation was the formation of nationalist soci­
11 Thus, to predict the international behavior of consolidated democracies, one
eties and pressure groups, whose repertoire otten featured rallies, lobbying, intimida­
should look to the theory of the democratic peace amon§' mature democracies, not to
tion, and other nonelectoral methods. These operated in democracies, such as the
the hypotheses about the war-proneness of democratizing states contained in this book.
28 From Voting to Violence
Most of the stares undergoing bloody ethnic conflicts that dominated
the news of the 1990s experienced a partial improvement in their politi­
cal or civil liberties in the year or so before the strife broke out. Most of
these conflicts occurred in states that were taking initial steps toward a
democratic transition, such as holding contested elections and allowing a
variety of political groups to criticize the government and each other.
Freedom House, an independent research and advocacy organization,
makes widely used annual rankings of every state's degree of democracy
and civil liberties, including press freedoms. By these measures, partial
democratization and partial increases in press freedom occurred before
the outbreak of ethnic conflict in the former Yugoslavia, before the esca­
lation in the fighting between Armenians and the former Soviet republic
of Azerbaijan, and in Russia, the perpetrator of the war against the eth­
nic separatist Chechens. In Burundi, the ethnic minority Tutsi military
government agreed to accept elections in 1993 and to share governmen­
tal offices with the long-oppressed Hutu majority. When, inevitably, the
elections installed a Hutu President who tried to insert some of his
coethnics in the military establishment, a conflict spiral touched off by
mutual fears and retaliation left some 50,000 dead. Though neighboring
Rwanda held no election, its Hum-dominated government allowed an
increase in press freedom on the eve of the 1994 genocide against the
Tursi minority.32
A systematic study of every ethnic conflict during the period from
1990 to 1998 noted that ethnic political assertiveness peaked during the
wave of transitions toward democracy that followed the collapse of the
Soviet empire from 1989 to 1991. 33 By mid-decade, this tide of democra­
tization had slackened or in some cases reversed itself, and correspond­
ingly the number of ethnic conflicts also diminished. 3' Where democracy
was successfully consolidated, as in much uf South America and the
northern part of Eastern Europe, minority rights were increasingly being
guaranteed through peaceful means. Meanwhile, fewer countries were
entering the dangerous category of semidemocratic regimes, where
31 Freedom House, Freedom in the World (New York: Greenwood Press, annual
yearbooks for 1989-94).
"Gurr, Peoples vmus Statf!S (forthcoming).
,.. Larry Diamond, "Is the Third Wave Over?" Journal of Democracy 7;3 Guly
1996),96.
Transitions to Democracy and thf! Rise ofNationalist Conflict 29
minority rights were more likely to suffer than to improve. Thus, democ­
ratic consolidation reduced ethnic conflict. but the initial steps in the
rocky transition to democracy increased it, especially in new states. 35
Democratic transitions have also tended to coincide with involvement
in international wars over the past two centuries. The chance of war in
any given decade for the average state has been about one in six, whereas
for democratizing states it has been about one in four during the decade
following democratization. The democratizing states were more likely to
be the attackers than the target of aggression in these wars. The most war­
prone states are those at the beginning stages of democratization, rather
than those that have nearly completed the consolidation of democracy.
Especially at risk for war are newly democratizing countries that lack a
strong centralized state to lay down firm rules for regulating popular par­
ticipation in politics and for enforcing state authority. 36 Similarly, domes­
tic political opposition to free trade rises significantly at the beginning of
democratization, whereas successful consolidation of democracy tends to
make states free-trading. Thus, the rise in nationalistic and uncooperative
thinking affects economic relations as well as military mattersY
This pattern of war-proneness during the early stages of democratiza­
tion echoes the history of virtually every great power. France, Britain,
Germany, and Japan all fought aggressive wars, fueled in part by popular
nationalism, on the heels of their initial phase of democratization.
"Gurr. Peoples vmus Statf!S (forthcoming).
,. See three studies by Mansfield and Snyder, "Democracization and the Danger of
War," "Democracization and War," and "Democratic Transitions, Institutional
Strength, and War." Although the darab3.!les used for these studies did not measure
nationalism directly, we presented logical argumenrs and narrative historical evidence
suggesting that nacionalism was at least one of the facmrs causing this link. On the
failure to consolidate and the heightened incidence of war, see Alexander
Kozhemiakin, Expanding the Zonf! of Peace? Democratization and InternationaJ
Security (New York: St. Martin's, 1998), Chaprer 5. For studies showing variations in
the strength of the link between democratization and war, see William Thompson
and Richard Tucker, "A Tale ofTwo Democratic Peace Critiques," Journal ofConflia
Resolution 41;3 Gune 1997), 428-~4. and other articles in that issue; John Oneal and
Bruce Russerr, "Exploring the Liberal Peace,» paper presenred at the Imernational
Studies Associacion meetings, April 1996, revised and shortened as "The Classical
Liberals Were Right," International Studitrs Quarterly 41 (1997),267-94.
31 Edward D. Mansfield, "Democratization and Comm~ial Liberalization," paper
presenred at the 1996 annual meeting of the American Political Science Association.
 Transitions to Democracy and the J?jse ofNationaList Conflict 31
30 From Voting to Violence

turned sharply in favor of forceful imperial expansion on the heels of

By today's standards, France was not a full-fledged democracy in the
the 1929 economic depression and the Japanese military invasion of
years immediately following the French Revolution of 1789: it had no
Manchuria in 1931.40 Thus, the turbulent early stages of democratization
stable rule of law, elections were irregular, and only some Frenchmen
in Germany and Japan were closely tied to the origins of both world wars.
could vote. Nonetheless, the elected assembly was the focal point of the
Although these episodes may seem like ancient history, Argentinas
nation's politics, and the press was, at least for a time, free and
fated 1982 invasion of the Falkland Islands, owned by Great Britain, was
Under these free-wheeling conditions, one of the revolutionary factions,
spurred by a similar dynamic. The Argentine military regime, its popularity
led by the newspaperman Jacques-Pierre Brissot, discovered that the best
waning, had recendy allowed more freedom of the press, which was in turn
way to win popularity and power was to hype the foreign threat. Brissot
effectively exploited by nationalist voices clamoring lOr a seizure of the
was swept into power through his allegations of a conspiracy linking the
islands. Hoping that a successful military gambit would enhance its popu­
Austrian monarchy to the French king and aristocracy. Soon the assembly
larity and position it well fur democratic elections that increasingly seemed
and the Paris political clubs (which nowadays we would call "civil soci­
inevitable, the Argentine military dictatorship gambled on attacking the
ety") were demanding war. The troops of revolutionary France marched
sovereign territory of a nuclear-armed great power and NATO member.';1
to the frontier in the name of liberty, equality, and fraternity, as well as
Thus, there are strong indications that nascent democratization and its
national defense. 38
close cousin, press liberalization, heighten the risk of nationalist and eth­
In comparison, British politics during the decades of partial democra­
nic conflict in our own time, just as they have historically. What accounts
tization following the First Reform Bill of 1832 were far less volatile.
for this correlation?
Nonetheless, even here, public opinion among the newly enfranchised
middle classes, inflamed by the nationalistic press, pushed the reluctant
British cabinet into the Crimean War against Russia in 1853. 39
Why Democratization Increases the Risk of Nationalist Conffict
World Wars I and II can similarly be attributed in part to the imper­
fect beginning stages of German and Japanese democratization. Before
Two contending views, which I label the "popular-rivalries" and the "elite­
1914, Germany had universal suffrage, an elected legislature that con­
persuasion" arguments, offer opposite explanations for the correlation
trolled the national budget, and habitual voter turnouts above 90 percent.
between democratization and nationalist conflict. The former contends
However, it was not a full democracy, since the hereditary monarch,
that long-standing popular nationalist rivalries precede democratization.
rather than the voters, picked the ministers who ran the executive branch
In this view, democratization gives expression to the long-held, popular
of government. Over a million German voters were members of national­
aspirations of an already-lOrmed nation, which are incompatible with the
ist organizations like the Navy League, which demanded an aggressive
aspirations of other nations. The "ancient hatreds" argument is one form
policy of imperial expansion that led Germany into confrontation with
of this popular-rivalries explanation.
Europe's great powers. Despite Germany's defeat in World War I, middle­
class voters once again helped install a belligerent nationalist government
Snyder, Myths ofEmpire (Ithaca: Cornell University Press, 1991), Chapters:
in the waning days of the unstable Weimar Republic, voting heavily for
3-4; Louise Young, "War Fever: Imperial Jingoism and the Mass Media,» in Japan'>
the Nazis in 1932. Japan, too, experimented with electoral democracy in Total. Empire: Manchuria and the Culture ofWartime Imperialism (Berkeley: University'
the 1920s. Japanese public opinion, encouraged by a relatively free press, of California, 1998).
" Richard Ned Lebow, "Miscalculation in the South Adantic,» in Robert Jervis, et:
al., Psychology and Deterrence (Baltimore: Johns Hopkins University Press, 1985)p
,. T. C. W. Blanning, The Origins of the French RevoLutionary Wan (London:
98-99; Jack Levy and Lily Vakili, "Diversionary Action by Authoritarian Regimes,'P
Longman, 1986).
in Manus Midlarsky, ed., The Internationalization of CommunaL Strife (London:
l' Norman Rich, Wkr the Crimtan Wan' (Hanover, NH: University Press of New
Rourledge. 1992), 118-46. •
England,1985).
):1. From I11ting to Violence
argue the opposite. Before democratization begins, nationalism is
usually weak or absent among the broad masses of the population.
Popular nationalism rypically arises during the earliest stages of democra­
tization, when elites use nationalist appeals to compete for popular sup­
pon. 42 Democratization produces nationalism when powerful groups
within the nation not only need to harness popular energies to the tasks
of war and economic development, but they also want to avoid surrender­
ing real political authority to the average citizen. For those elites, nation­
alism is a convenient doctrine that justifies a partial form of democracy,
in which an elite rules in the name of the nation yet may not be fully
accountable to its people. Under conditions of partial democratization,
elites can often use their control over the levers of government, the econ­
omy, and the mass media to promote nationalist ideas, and thus set the
agenda for debate. Nationalist conflicts arise as a by-product of elites'
efforts to persuade the people to accept divisive nationalist ideas.
It maners which of these two views is correct, because each points
toward different prescriptions for averting nationalist conflict. If the pop­
ular-rivalries view is right. the preferred solution should often be to parti­
tion democratizing ethnic groups into separate states, even if it means
moving populations. 43 Much of Eastern Europe has already undergone a
vast "unmixing of peoples" in the twentieth century, most of it as a result
of extermination and forced emigration during and immediately after the
two world wars. 44 Why not, in this view, move people before the fighting
rather than afterward? Where partition of a multiethnic state is impracti­
cal. the closest substitute would have to be adopted: powersharing
between largely self-governing nationalities. Under this scheme, some­
times called "consociational democracy," people would have rights not
as individuals but also as members of a national or ethnic group,
" For related arguments, see V. P. Gagnon. "Ethnic Nationalism and International
Conflict: The Case of Serbia," International Security 19:3 (Wmter 1994/95), 130-66;
V. P. Gagnon. "Ethnic Conflict as Demobilizer: The Case of Serbia' (Cornell
University. Institute for European Studies Working Paper no. 96.1, May 1996); Paul
Brass, Language. &ligion and Politics in North India (Cambridge: Cambridge
University Press, 1974).
<3 For a discussion of the conditions in which partition is necessaty, see Chaim
Kaufinann, "Possible and Impossible Solutions to Ethnic Civil Wars," International
Security 20:4 (Spring 1996), 136-75.
44 Rogers Brubaker. Nationalism &framed, Chapter 6.
Transitions to Democracy and the Rise ofNationalist Conflict 33
which would govern its own intemal affairs and have group rights to pro­
ponional representation in the bureaucracy and the legislature. 45
However, if the elite-persuasion view is correct, such separation measures
might serve to lock in divisive national identities, unnecessarily heighten­
ing distrust between groups. When this is the case, a better solution
would be to take advantage of the fluidity of national identity during the
formative stages of democratization to promote more inclusive, civic
identities and cross-ethnic political alignments.
I will first summarize and assess the popular-rivalries view, in part
because it seems so plausible, indeed commonsensical, to most American
readers. Then I will briefly summarize my own elite-persuasion argument,
which will be presented more fully in Chapter 2.
PreJemocratic popular rivalries:
A simple but usually errtJft!eO'IU explanation
A very simple explanation for the link between democratization and
nationalist conflict sees nationalism as deeply rooted in popular attitudes
that long precede democratization. If people of different cultures natu.­
rally want their own state, and if they inhabit the same territory, giving
them the vote will unavoidably put them at loggerheads. In principle,
these conflicting aims might cause conflict even if the contending groups
had no predemocratic history of violence toward each other. However, if
people of different cultures had learned through long-term rivalries to
hate or distrust each other, that would make the rivalry even worse; the
democratic process would express those entrenched popular animosities.4<i
In the popular-rivalries view, national strivings can be suppressed by
empires and authoritarian regimes, whereas democratization gives voice
to the true nationalist preferences of the average voter. Elections become a
census rather than a deliberative process. Democratization will tend to
" Arend Lijpharr, Democracy in Plural Societies (New Haven: Yale University Press,
1977).
<6 Hurst Hannum, Autonom)\ Sovereignty. and Self-Determination (Philadelphia:
University of Pennsylvania. 1996), discLL';ses problems arising OUt ofdemands for eth­
nic self-determination; Daniel Byman and Stephen Van Evera, "Hypotheses on me
Causes of Contemporary Deadly Conflict," Security Studies 7:3 (Spring 1998), esp.
33-35, discuss the conflictual consequences of the "ryrarmy of the majority" and of
historic grievances of "hardened groups" in democratizlng'mulriethnic societies.
34 From ~ting to Violence
produce either the tyranny of the majority or a pitched battle between the
competing state-building goals of rival nations. Democratization, in this
view, may also give rise to international conflicts, if a newly democratic
nationalizing stare seeks to capture foreign territory where its ethnic
brethren reside.
Many Americans, steeped in Woodrow Wtlsons doctrine of national
self-determination, take for granted that humanity is divided into distinct
peoples, each of whom has the natural desire to rule itself in its own
way.47 When the Soviet Union collapsed, for example, many Americans
with Wtlsonian instincts saw it as quite natural that each of the fifteen
ethnically tided constituent republics of the USSR should exercise its
democratic right to self-determination. This seemed all the more just
because of the long history of Soviet repression of what U.S. cold warriors
called the "captive nations": mass deponations of some non-Russian eth­
nic groups to Siberia and Central Asia, systematic campaigns of terror or
starvation againSt others, and dictatorial rule from Moscow over all the
subject nations at the periphery of the empire. Emerging from this his­
tory, the formerly captive nations had ample reason to desire self-rule
rather than rule by Moscow.
But where ethnic groups live side by side in the same towns and
regions, one group's aim to establish self-rule in its own state is sure to be
incompatible with that of others.48 Every successor state to the Soviet
Union except Armenia houses a substantial minority, including the
Russians who populated many of the cities and industrial enterprises of
the non-Russian Soviet republics:9 Under these circumstances, democra­
tic self-determination by ethnic groups becomes a recipe for conflict if
each group tries to establish a state in which it monopolizes citizenship
rights and in which policies on linguistic or economic issues seIYe the
interests of the dominant ethnic group. Because so few states in the for­
mer Soviet empire or in the developing world are ethnically homoge­
41 For a critique of this view, see David Laitin. Idmtity in Formation (Ithaca:
Cornell University Press, 1998), 12.
4. Hurst Hannum, "Rethinking Self-Determinacion," Virginia Journal of
Internatio1llZi Law 43:1 (Fall 1993), 1-69, poinrs out the dusiveness of sdf-determi­
nation doctrine on the question of the boundaries of the sdf-determining unit.
49 Robert Kaiser, The Geography of Natio1lJZ/i.sm in Rmsia and the USSR (Princeton:
Princeton University Press, .994). Armenia, which had the smallest minority population
even in Soviet times, expelled most ofits Azerbaijani minority before independence.
Transitions to Democracy and the Rise ofNationalist Conflict 35
neous, Wtlsonians should not be surprised if many new democracies are
conflict-prone. 50
Despite the dear logic behind this popular-rivalries view, I argue that
in most cases it gets the faCts exacdy backward. Mass nationalism is rarely
well developed before democratization. More conunonly, it rises during
the earliest phase of democratic change. In the era before the majority of
the population takes an active part in political life, their sense of belong­
ing to a nation is usually weak.. Typically, they ate aware of cultural, lin­
guistic, religious, and regional differences across groups, but they attach
political significance to these differences only intermittendy, if at all. At
that stage of social development, politics is a matter for elites. It is true
that cultural legacies or administrative arrangements based on nationality
in the authoritarian regime may load the dice in favor of the fater devel­
opment of a particular form of nationalism, once the broader population
becomes politically active. Even so, in most cases, this consciousness does
not crystallize until people start to speak out in public and form mass
political organizations along national lines.
For example, it would be wrong to view Serbias war for autonomy
from the Ottoman Empire in the early nineteenth century as an outpour­
ing of ancient popular nationalist sentiment. As I show in Chapter 4, this
struggle was more a commercial enterprise by a multiethnic cabal of pig­
traders than a war of national liberation. Serbian national consciousness
took shape only gradually from the 1860s to 1914, when Serbia had a
surprisingly democratic though disorderly political system with nearly
universal suffrage, competing panies, and a free press. The wars and
hatreds of that period made a lasting impact on the national conscious­
ness, in part because the public was already playing a role in political life.
As in most of the European states, according to historian Miroslav
Hroch, "the process of nation-forming acquired an irreversible character
only once the national movement had won mass suppon. "51
50 Clifford Geertz, "Primordial Sentiments and Civil Politics in the New Scates,"
in Geerl.'Z, ed., Old Societies and New States (New York: Free Press of Glencoe, 1963),
123. According to one survey. fewer than 10 percent of scates are populated almOSt
entirdy by members of a single national group. Walker Connor. Ethno1llZtionaJism
(Princeton: Princeton University Press, 1994), 96.
5l Miroslav Hroch, "Real and Constructed: The Nature of the Nation," in John
Hall, ed., The State of the Nation: Ernest Gellner and thi Theory of Nati01llZiism
(Cambridge: Cambridge University Press, 1998).
36 From Voting to Violence
In short, nations are not simply freed or awakened by democratization;
they are formed by the experiences they undergo during that process. The
type of political experiences, institutions, and leadership that prevails dur­
ing the initial phases of democratization can be decisive for the formation
of national identity. How people are included in the political life of their
state determines the kind of national consciousness that they develop, as
well as the degree of nationalist conflict that democratization brings.
Consequently, it is of immense practical importance, as well as academic
interest, to explore in detail what forces shape the kind of nationalism
that emerges from the crucible of democratization. PurpOrted solutions to
ethnic conflict that take predemocratic identities as fixed, such as parti­
tion, ethnofederalism, ethnic powersharing, and the granting of group
rights, may needlessly lock in mutually exclusive, inimical national identi­
ties. In contrast, creating an institutional setring for democratization that
de-emphasizes ethnicity might turn these identities toward more inclu­
sive, civic self-conceptions.
Elite perstUUion: Selling nationalism in democratizing states
Democratization gives rise to nationalism because it serves the interests of
powerful groups within the nation who seek to harness popular energies
to the tasks of war and economic development without surrendering real
political authority to the average citizen. In predemocratic societies, mili­
tary, economic, and cultural elites preferred to rule without taking the
risk of arousing popular nationalist sentiments. However, with the rise of
mass armies, commercial capitalism, and inexpensive printing in the eigh­
teenth and nineteenth centuries, rulers who could attract the active sup­
port or their people gained a valuable advantage in competition against
other states and against rival elites at home. In a growing number of
countries, old elites either granted some democratic reforms, or else rising
new elites forced such changes. Despite the pressures for democratization,
both the old and new dites were typically reluctant to allow full democra­
tic rights, since this could have endangered their parochial economic
interests as well as their positions of power in society. Nationalism, a doc­
trine of rule in the name of the people but not necessarily by the people,
provided a way for elites to be popular without being fully democratic.
A very effective tool for containing popular pressure for democratiza­
tion is the use of nationalist doctrine to exclude so-called enemies of the
Transitions to Democracy and the Rise ofNationalist Conflict 37
nation from enjoying democratic rights. Nationalist elites commonly
argue that ethnic minorities, the working classes, rival elites, or other
political opponents should be excluded from political participation, often
alleging that these groups lack the proper national credentials and are in
league with foreign powers. This tactic not only justifies curtailing the
democratic rights of these purported "enemies within," but just as impor­
tant, it has a chilling effect on freedom of expression among all citizens. 52
The prospects for elite lrtempts at nationalist persuasion depend in
part on the timing of democratization relative to the development of the
country's economy and political institutions. Exclusionary nationalism is
most likely to prevail when the democratizing country is poor, when its
citizens lack the skills needed for successful democratic political participa­
and when its representative institutions, political parties, and jour­
nalistic professionalism are weakly established during the early phase of
the democratic transition. In these conditions, nationalist elites are more
able to hijack political discourse. Moreover, in such a barren and
unpromising political landscape, threatened elites are likely to have little
confidence that a fully democratic regime could reliably guarantee to p .£0­
tect their interests after they surrender power. Under these conditions,
gambling on staying in power at the crest of a nationalist tide will seem
an artractive alternative. Conversely, exclusionary nationalism is less likely
to thrive in countries like nineteenth-century Britain or contemporary
South Africa that democratize after the necessary economic resources, cit­
izenship skills, or political institutions are already in place.
These outcomes depend on both the motivation and the opportunity of
elites to promote nationalist doctrines. The strength of an elite's motivation
depends on the adaptability of its interests to a more democratic setting.
The more the elite feels threatened by the arrival of full democracy, me
stronger is its incentive to use nationalist persuasion to forestall that out­
come. The elite's opportunity to sell exclusionary nationalism depends to a
large extent on the character of the political institutions of the democrariz­
ing state. For example, where the state's bureaucracy is strong, yet its insti­
tutions for democratic participation and public debate are weak, scate
elites will be able to use their administrative leverage to promote national­
12 Gagnon, "Ethnic Conflict as Demobilizer"; on the role of racial exclusion, see
Anthony Marx, Making Race and Nation (Cambridge: Cambridge University Press,
1"~ .
 From Voting to Violence
38
ism during the early phase of democratization. Conversely, when represen­
tative and journalistic institutions are already well developed during early
democratization, nationalist arguments can be checked through more
effective scrutiny in open public debate.
Together these two factors, the adaptability of elite interests and the
strength of the country's political institutions during early democratiza­
tion, determine the intensity of the democratizing country's nationalism
and the form that nationalist exclusions are likely to take. Representing
the resulting possibilities as a schematic simplification, this yields four
types of nationalism: counterrevolutionary, revolutionary, ethnic, and
civic (see Table 1.1). Three of these four entail severe exclusions of differ­
ent kinds and are likely to lead to intense nationalist conflicts.
Counterrevolutionary nationalism is likely to emerge when elite inter­
ests are not adaptable, and when administrative institutions are strong,
but representative institutions are weak. In that situation, my theory pre­
dicts that attempts at nationalist persuasion will be intense and etIective.
Threatened ruling elites will justify excluding political opponents from
power by portraying them as revolutionary enemies of the nation. As in
Germany before World War 1, nationalism will take a counterrevolution­
ary form.
Conversely, revolutionary nationalism will emerge when state institu­
tions have already collapsed, and opportunistically adaptable elites seek to
establish a popular basis for restoring power to the state. Under those con­
ditions, I predict that nationalist persuasion will be used etIectively to rally
support against foes at home and abroad. As in the French Revolution,
nationalism will take an intensely exclusionary revolutionary form.
Ethnic nationalism is likely when democratization begins in a setting
where the basic building blocks of political or administrative institutions
have never been laid down. In this institutional desert, elites will by
default be constrained to base appeals to loyalty on the only available
alternative, traditional popular culture. As in nineteenth-century Serbia,
an intensely exclusionary nationalism will take an ethnic form.
These three types of exclusionary nationalism--counterrevolutionary,
revolutionary, and ethnic-are likely to produce violent nationalist con­
flicts with the excluded groups inside the country and with any of these
groups' purported foreign allies. The fourth variant of nationalism, the
civic type, is more moderate and inclusive.
Nationalism will take an inclusive civic form when elites are not partic-
Transitions to Democracy and the Rise ofNationalist Conflict 39
ularly threatened by democratization, and when representative and jour­
nalistic institutions are already well established before the mass of the
population gains political power. Under those conditions, as in Britain in
the eighteenth and nineteenth centuries, nationalists lack both the motive
and the opportunity to purvey divisive doctrines. While civic nation­
alisms are not predicted to be pacifistic, they have far less reason to fail
prey to the kind of reckless, ideologically driven conflicts characteristic of
the other three types.
Table 1.1 Relationship ofPoliticallnstitutioDS and FJjtes' Interests to the
Type of Nationalism during the Early Phase of Democratization.
Strength ofthe Nation's PolitielllIrutitutions
NationAlist Elite!l'
Interests Strong weak
Adaptable Civic: strong representative Revolutionary
institutions (Britain) (revolutionary
France)
Unadaptable Counterrevolutionary: Ethnic (pre-World
strong administrative War I Serbia)
institutions (pre-World
War I Germany)
Making Choices in Todays World
These explanations for the connection between democratization and
nationalism have significant implications for designing strategies to man­
age contemporary nationalism. The ethnic conflicts of the 1990s spurred
a number of sharp policy debates among politicians, journalists, and
scholars. Understanding how democratization causes nationalist conflict
sheds new light on each of those debates.
The most general debate revolves around the nature of the coming
epoch, whether it will be wracked with ethnic conflict, or whether it is
heading toward an inclusive democratic peace. The analysis advanced in
this book suggests that both perspectives are likely to be right. The democ­
ratic peace will prevail within and between states wherever democracy is
quickly and successfully consolidated. Meanwhile, nationalism will be on
40 From voting to Violence
the rise in the many states that find themselves newly embarking on
democtatization, stuck in betWeen autocracy and democracy, suffering
from the consequences of failed democratization, or simply anticipating its
unsettling effects. Insofur as great powers like China and Russia may face
these dangers, dealing with democratization as a cause of nationalist con­
flict should remain high on the agenda of the international community.
More concretely, the management of these transitions provokes sharp
differences of opinion. Some argue that international organizations, non­
governmental human rights groups, and powerful democratic states ought
to press all dictators to democratize immediately. New York Times editorials
routinely insist that the authoritarian leaders of such countries as Congo
and Malaysia should immediately announce multiparty elections. My the­
ory suggests skepticism about that advice. Given the postcolonial pattern
of patronage and political alignment in such countries, factionalism in pol­
itics during their democratization often follows ethnic linesY And given
that starting point, effective institutions for channeling social cleavages in
other directions need to be well developed before democratization can be
part of the solution rather than part of the problem.
Some argue that the international community should press for power­
sharing betWeen groups in multiethnic societies, with ethnic groups, not
just individuals, enjoying rights. 54 My theory suggestS that this is a last
resort because it runs the risk of unnecessarily politicizing and locking in
inimical cultural distinctions. Wherever possible. democratizing states
should try to promote civic identities and guarantee rights at the individ­
ualleveL For the same reasons, ethnically based federalism and regional
autonomy should be avoided, since they create political organizations and
media markets that are centered on ethnic differences.
Some argue that multiethnic states with a history of intercommunal
conflict ought to be partitioned before the conflict recurs, even if this
means moving people to new homes. 55 My theory suggests that partitions
might be necessary where ethnic conflicts during formative periods of
Mahmoud Mamdani, Citizen and Subject: Contemporary Africa and the Legacy
Il
LAte Colonialism (Princeton: Princeton University Press. 1996), 300.
54 Arend Lijphan, "The Power-Sharing Approach,n in Joseph V. Montville, ed.,
Conflict and Peacemaking in Afultiethnic Societies (New York: Lexington, 1991),
491-510.
lS Chaim Kaufmann, «Possible and Impossible Solutions to Ethnic Civil Wars,"
international Security 20:4 (Spring 1996), 136-75.
Transitions to Democracy and the Rise ofNationalist Conflict 4I
democratic development have already created entrenched institutions,
ideas, and interests based on invidious ethnic distinctions. But for states
that have not undergone significant levels of democratization, such parti­
tions should be unnecessary if leaders are willing to adopt a strategy of
civic institution-building before embarking on democratization. The grad­
ual development of the rule of law, an impartial bureaucracy, civil rights,
and a professional media, followed by the holding of free elections, should
be able to create a civic national identity that trumps "ancient hatreds."
Many human rights organizations argue that individuals responsible
for human rights abuses in waning authoritarian systems, especially their
leaders, ought to be punished for their crimes against humanity. My the­
ory, however, highlights the incentive that this creates for still-powerful
elites to play the ethnic card in a last-gasp effort to forestall this fate.
Indeed, democratization has proceeded most smoothly where authoritar­
ian elites were given a "golden parachute" into a safe retirement, as in
South America and northeast Europe. Where elites have felt most threat­
ened, notably in Rwanda. human rights disasters have only intensified.
Punishment is a prudent strategy only when the human rights abusets are
too weak to wreak such havoc.
Finally, advocacy groups like Human Rights Watch propose great:er
freedom of speech and a more active "civil society" as antidotes to
manipulative governments that foment ethnic conflict. 56 Yet in many
infant democracies, the newly fteed press becomes a vehicle for nation­
alist appeals. Weak democratic institutions often make society uncivilY
Unfettered speech and a vibrant civil society are forces for peace only
when conditions are favorable, that is, when media audiences are
sophisticated and journalists are professionalized. Projects to foster those
preconditions need to precede proposals to unleash vigorous debate in a
free press.
These are controversial prescriptions that run counter to the current
conventional wisdoms of the international human rights community as
I. Human Watch. Slaughter among Neighbors (New Haven: Yale University
Press, 1995).
57 Sheri Berman, "Civil Society and rhe Collapse of me Weimar Republic," Wo,./d
Politics 49:3 (April 1997), 401-29; Margaret Levi, "Social and Unsocial Capital."
Politics and Society 24 (March 1996), 45-56; Sidney Tarrow, "Making Social Science
Work across Time and Space:' American Political Science, Review, 90:2 Oune 1996),
389-98.

From Voting to Violence
42
well as many of the natural instincts of the American public. Nonetheless,
I believe that the theoretical logic and case-study evidence presented in
the coming chapters should make even the most avid proponent of
spreading democracy and human rights think more carefully about devis­
ing ptudent ways and means of pursuing those ends.
The Plan of the Book
The next chapter fleshes out my elite-persuasion explanation for the cor­
Transitions to Democracy and the Rise ofNationalist Conflict 43
well each state's distinctive pattern of democratization can explain the
degree of its nationalism.
Chapter 6 discusses the relationship between democratization and
nationalism in the contemporary developing world. It begins with a
broad overview of the conditions that have led to intense nationalist and
ethnic conflicts in some democratizating states, and contrasts them with
the conditions that have led to peaceful transitions in others. Then I look
in more detail at a few cases that illuminate key issues for my theory. A
comparison of the closely parallel cases of Sri Lanka and Malaysia shows
how unfettered democratization triggered ethnic strife in the former case,

whereas in similar circumstances, a scheme imposed by an authoritarian

relation between democratization and nationalist conflict. It also discusses

leader averted ethnic conflict in the larter one. Sketches of Rwanda and
some possible alternative explanations. To probe the plausibility of these
Burundi show the dangers of liberalization in states lacking an adequate
arguments, I examine historical as well as contemporary cases of democ­

institutional foundation, while the case 0f India illustrates the conse­

ratization. I rely on a combination of twO strategies: tracing causal

quences that ensue when those foundations begin to atrophy.

processes within cases and comparing across cases. Chapter 7 draws out the practical implications of the analysis, assess­
Chapters 3 and 4 present historical case studies of the four types of
ing various prescriptions for averting nationalist conflict. Overall, I argue
nationalism: counterrevolutionary Germany before the twO world wars, that the international community should putsue a patient and sometimes
civic Britain in the eighteenth and nineteenth centuries, revolutionary
indirect strategy for promoting the spread of wdl-institutionalized civic
France, and ethnic Serbia in the century before 1914. These four cases democracies. Shortcuts on the road to the "democratic peace" could wind
have been focal points for much of the theoretical and historical literature
up as detours to a counterrevolutionary nationalist backlash or to rival­
on nationalism, and for good reason. Britain and France, the twO earliest rous ethnic pseudodemocracies.58
cases of nationalism, created a form of political loyalty that other states
felt compelled to emulate. France and Germany carried out nationalist
aggression on a continental scale. Serbia spans twO centuries as the quin­
tessential example of divisive ethnic loyalty. Any theory of nationalism
must come to grips with these seminal instances. I devote all of Chapter 3
to the German case, which provides an exceptionally rich laboratory illu­
minating many of the mechanisms through which democratization causes
nationalism and nationalist conflict. Chapter 4 sketches the other three
cases and concludes by comparing all four types of nationalism. These
paradigmatic cases illustrate patterns that help in assessing the contempo­
rary case studies presented in subsequent chapters.
Chapter 5 examines the impact of democratization on nationalism and
ethnic conflict in the post-Communist states. It is important to analyze
these cases both because of their intrinsic contemporary significance and
because they constitute a hard test for my theory. Many of the post­
58 For a related argument, see Pareed Zakaria, "The Rise of Illiberal Democracy, ff

Communist states democratized, yet not all of them experienced intense

Foreign Affairs 76:6 (NovemberlDecember 1997), 22-43.
nationalist conflicts. Thus, this set of cases presents a challenge to see hoW